TL;DR version: no, it has not.

Last night, the ABC unfortunately saw fit to give a platform to a spot of statistical scaremongering that has been floating around for a while: the idea that there’s something dodgy about the official unemployment figures. The piece, which ran on The Business, was based around the claims made by Andrew Baker from the CIS in a recent policy paper about the effect of some government policy changes a few years ago on the unemployment numbers. The ABC piece was titled “Lies, more lies, propaganda and the unemployment rate,” a fairly unsubtle suggestion that there’s something distinctly dodgy about the numbers.

There seems to be a rich vein of scepticism out there about the official figures, which I think ranges from the misguided and misinformed to the cynical and conspiratorial. As far as I can tell, there are three key reasons why people mistrust these numbers, which I’ll address in turn.

1) The unemployment rate is an incomplete and imperfect measure of the state of the labour market

This is obviously true. There are nearly 19 million civilians aged 15 or over in Australia who fall within the scope of the ABS labour force survey – no single measure could ever adequately summarise the range of ways they engage with the world of work. There are people who are working who would like more work, people who would like less work, people who are out of work and are quite happy about that, others who desperately seek paid work. There are people who have been stood down for a while, people on strike, people injured at work who are doing an hour or two here and there to get back into the workforce. How could one statistic ever hope to fully encapsulate the breadth of that experience? How could one measure tell us everything we need to know about how the demand for labour has changed relative to its potential supply?

The unemployment rate is inadequate, as is any other alternative measure you’d like to propose. I quite like looking at the employment-to-population ratio – this simple measure tells us what proportion of the population is in work. Ah ha, you might ask, but what if all the newly created jobs are part time? Good point – for that reason I also quite like looking at the number of hours worked per adult in the population. But this isn’t perfect either, as when hours fall it doesn’t tell us whether this pain has been shared broadly across the population, or concentrated in a few unlucky people who’ve gone from full-time work to the dole queues. If you want to get a sense of how the labour market is going, you’d want to look at all of the above measures and more.

I think this is what John Buchanan of the Workplace Research Centre was getting at in the story on The Business when he said “as a labour market researcher, the unemployment rate is one of the least helpful indicators for understanding what’s going on.” I think he meant that it tells us nothing about the nature of people’s experience at work, about how long the unemployed have been out of work, or about the number of people in work who would like more hours. Unfortunately, the sequencing of the story made it seem as if he was agreeing with the claims made by Andrew Baker; I would be surprised if that was the case.

The unemployment rate is imperfect. This is both true and trite.

2) The ABS unemployment rate doesn’t line up with the Roy Morgan figures

If you work for an hour during the week the ABS asks you about, then it classified you as employed. If you’re not in work, but you haven’t actively looked for work in the past four weeks or you’re not able to start work in the survey week, then you’re not ‘unemployed’ in the eyes of the ABS – you’re deemed to be not participating in the labour force.

Gary Morgan of Roy Morgan Research doesn’t agree with the way the ABS defines unemployment. He says the ABS numbers merely measure the ‘perception’ of unemployment, while his figures give the ‘real’ unemployment rate. I don’t think this is a fair characterisation.

The ABS didn’t make up the definitions it uses for ‘employed’ and ‘unemployed’. They’re internationally accepted definitions, adopted by successive international conferences of labour statisticians at the International Labour Organization, a tripartite UN agency that consists of government, business, and union representatives. This way of measuring employment and unemployment has been the international standard for several decades.

Morgan doesn’t like this standard. While the ABS will only classify a person as unemployed if you’ve actively looked for work sometime in the past four weeks, for Morgan it’s good enough “if they are looking for work, no matter when.” So if you wouldn’t mind a job, but you haven’t actually hit up for a couple of months, no matter – in Morgan’s eyes you’re just as unemployed as the desperate person who’s combing the Centrelink jobs board every day.

Fair enough, but by using this definition Morgan is measuring something different than the ABS and most of the national statistical agencies of the world. They’re measuring oranges, Morgan is measuring all citrus fruits. It’s a related measure, but clearly broader. There’s no reason to expect that measuring two different things will give you the same result, or even that the figures will necessarily move in the same direction at the same time. The ABS survey around 26 000 households every month and around 97% of them respond – not surprising, given that the ABS can legally compel you to do so. I don’t know how many people Roy Morgan surveys and I don’t know how many respond, but I’d be pretty hesitant to believe either figure is higher than the ABS equivalent.

If you think that the underemployed are important, – those people who have a job, but would like more work – I agree with you. But there’s no conspiracy of silence around underemployment. Around 900 000 people fit into this category in May 2013 – 7.4% of the labour force. Add that to the unemployment rate and you have the ‘labour force underutilisation rate’ of 12.9%. I don’t know these numbers because some other pollster uncovered the truth. They’re right there in table 22 of the official figures.

3) The official unemployment rate has been manipulated

First, a bit of background. The unemployment rate that politicians and economists refer to comes from the ABS labour force survey, as I mentioned above. The ABS asks people a number of questions about their work or lack thereof and puts them in a category based on their answer. The criteria the ABS uses don’t ask about whether you’re receiving a government benefit or not. You could be on the dole and not considered unemployed by the ABS (because you’re working part time, or not looking for work because you’re in training, for example). You could be ineligible for Newstart (for example, if you have too much cash in the bank), but still fit the bill as far as the ABS/ILO criteria are concerned. The official unemployment figures aren’t based on the number of people who are claiming benefits from Centrelink. The ABS unemployment numbers and the benefit claimant figures are like an orange and a mandarin – they’re clearly related, but not quite the same.

You might have got another impression if you’d watched that segment on The Business. In it, Andrew Baker says “the government has been manipulating the unemployment rate to reduce it, for their political benefit.” Big call. Bob Gregory, a very well-respected ANU economist, says “I just think that’s a ridiculous proposition. That’s just not true.” By this point it will not surprise you to learn that I agree with Professor Gregory.

Baker’s claim of manipulation rests on a blurring of the lines between the ABS definition of unemployment and the number of jobseekers claiming Newstart Allowance. He points out that in July 2009, some benefit rules were changed – a slightly larger proportion of people on Newstart now can meet the requirements of the payment by training, by developing a small business, or by undertaking part-time work. While Baker’s research paper does note that the ABS figures and the claimaint numbers aren’t the same thing, you could easily get a different impression from the story on The Business.

I just don’t buy Baker’s argument. Have a look at the chart below. The number of people on Newstart (the orange line) has tracked pretty closely with the number of people the ABS thinks are unemployed. There’s no obvious change from July 2009. That jump up in January 2013 happened because we kicked a bunch of single parents off the parenting payment and onto Newstart.

Screen Shot 2013-08-29 at 10.30.28 PM

The idea that some modified rules in the benefit system amount to manipulation of the unemployment rate is a “ridiculous proposition”, as Bob Gregory put it. It’s a shame the ABC so uncritically accepted this point of view, and that it has been reiterated elsewhere.

The boring truth is that the unemployment rate is an imperfect, but very useful measure of an important, widely accepted aspect of the labour market. I am quite sure that there has been no conspiracy to manipulate the unemployment rate in Australia.